JDevlieghere wrote: > We agreed that the native parser would be the way forward in the last > EuroLLVM round-table IIRC (CC @JDevlieghere @labath).
That's right. The problem is that neither implementation is complete, and things only work because there's an automatic fallback. No matter which one you pick, it's going to introduce some regressions. Conceptually, everyone is in agreement that the native parser is the way forward: - It means we're not limited to testing PDF support on Windows hosts. - It means we have the ability to improve support (we don't have control over DIA). > I think Jonas had a PR open for it but there were some small test blockers. > What was the outcome of that work @JDevlieghere ? Here's the link to that PR: https://github.com/llvm/llvm-project/pull/113647. At the time, removing the DIA implementation caused 72 test failures. Shortly after, @ZequanWu put up some PRs that improved the situation, but I haven't rebased my PR since. Depending on the number of failures, we can reopen the discussion. I totally understand that folks relying on the DIA implementation are hesitant to regress, but I believe it's in the best interest of LLDB. https://github.com/llvm/llvm-project/pull/149305 _______________________________________________ lldb-commits mailing list lldb-commits@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/lldb-commits
